Pitch characteristics of real-world infant-directed speech vary with pragmatic context, perceived adult gender, and infant gender.

Children's everyday language environments can be full of rich and diverse input, especially adult speech. Prosodic modifications when adults speak to infants are observed cross-culturally and are believed to enhance infant learning and emotion. However, factors such as what and why adults are speaking as well as speaker gender can affect the prosody of adults' speech. This study asks whether prosodic modifications to infant-directed speech depend on perceived adult speaker gender, assigned infant gender, and the perceived pragmatic function of an utterance. We examined 3,607 adult speech clips from daylong home audio recordings of 60 North American, English-speaking, 3- to 20-month-old infants (28 female). Adult speakers used significantly more imperatives and questions and sang more frequently to infants than other adults. While infant-directed speech tended to have greater mean pitch and pitch modulation than adult-directed speech overall, these patterns were modulated, sometimes in complex ways, by pragmatic function, perceived adult gender, and infant gender. For example, we found that female-sounding adult speakers exhibited greater IDS-ADS mean pitch differences than male-sounding adult speakers when providing information or engaging in conversational niceties. An additional example is that male-sounding adults used higher pitch when singing to male infants compared to female infants. These findings invite further research on how individual, demographic, and situational factors affect speech to infants and possibly infant learning. The study's pragmatic context tags are added to an existing open dataset of infant- and adult-directed speech.